Steve Bennett wrote:
Hi,
I didn't understand exactly what you're proposing. Can you clarify?
On 5/28/06, GerardM <gerard.meijssen(a)gmail.com> wrote:
Hoi,
You have to allow for the way the pywikipedia bot works. There is a
whole infrastructure around MediaWiki. and interpreting these things
differently depending will create breakage.
Interpreting which things differently? What is the piwikipedia bot?
The Pywikipedia bot is the (python) software that is used by many if not
most of the bot operators on MediaWiki projects. When the MediaWiki
software takes a different tack, the pywikipedia software has to follow.
What I am suggesting is that changing the behaviour of MediaWiki can
cause breakage in the bot environment (it did in the past).
Depending on what software you use two double
quotes and two two
apostrophes are displayed in a different way. It is therefore
Yes, I'm not necessarily suggesting every displaying a double quote as
if it was two apostrophes.
important to keep things as they are in that
language. It is imho much
better to have language dependent behaviour.
What is "language dependant behaviour" in this context exactly?
Steve
At this moment, a Wikimedia project is associated with a project code.
This code has a passing resemblance with language codes. At this moment,
for a project there is an association with language aspects; for
instance the Hebrew, Arab et al are left to right where we typically
expect them to be right to left. Gangleri, Nikerabbit and Rob Church
have spend quite a lot of effort in fixing many issues in MediaWiki to
deal with those. They are doing a stellar job making these necessary
fixes to the MediaWiki usability for the left to right scripts. These
are however primarily script dependent fixes. There are other issues
where there the fix is necessary on the level of the language.
There are also issues where there is a need for language dependant
behaviour; there will be a large improvement with the Multilingual
MediaWiki where it will be possible to associate an article with a
language. When we can associate the article with a language, the next
step would be to indicate a text within an article with a language. An
example how this can already be done can be found on the WiktionaryZ
main page where several of the language codes have been associated with
language dependant. Here you find for instance that the Assamese text is
shown in a different way from what is standard.
Yesterday there was a post of people translating English Wikipedia
content to Navaho. When we mark content that is decidedly NOT English,
it will make the machine translation engine more effective. I would
advocate identifying text to be identified as not English content when
we had the Multilingual MediaWiki, however at this moment it may be too
early to do this.
Thanks,
GerardM